Human Reproduction, Vol.30, No.11 pp. 2671– 2676, 2015
Advanced Access publication on September 9, 2015 doi:10.1093/humrep/dev200
ORIGINAL ARTICLE Reproductive epidemiology
A mass shooting at Port Arthur,
Tasmania, Australia: a study of its
impact on early pregnancy losses using a
conception time-based methodology
R.G. Dean1,*, J. Dean2, G.Z. Heller 3, and L.R. Leader 1
1
School of Women’s and Children’s Health, Medical Faculty, The University of New South Wales, Kensington, NSW 2052, Australia 2formerly,
AIHW National Perinatal Epidemiology and Statistics Unit, School of Women’s and Children’s Health, Medical Faculty, The University of
New South Wales, Kensington, NSW 2052, Australia (now retired) 3Department of Statistics, Macquarie University, Sydney, NSW, Australia
*Correspondence address. E-mail: [email protected]
Submitted on November 17, 2014; resubmitted on July 16, 2015; accepted on July 23, 2015
study question: Does an acute calamity in a community cause early miscarriage and is this association the same for male and female
fetuses?
summary answer: Estimated losses of 29.5% of first trimester pregnancies in the affected region could be associated with an acute
calamity, with no statistically significant difference in estimated losses by fetal sex.
what is known already: There are very few studies on the impact of a calamity on early pregnancy loss and its differential effects on
male and female fetuses. A decline in the human sex ratio at birth associated with the events of 9/11 in New York has been documented.
study design, size, duration: This is a retrospective descriptive study of birth register data in Tasmania, Australia, from 1991 to
1997, covering the period in which the calamity occurred. The register contains data on all pregnancies that proceeded to .20 weeks gestation.
The conception date was calculated by subtracting gestational age from birth date. We estimated that 40 318 pregnancies were conceived in the
period 1991– 1996 inclusive. These were aggregated to 4-weekly blocks classified by region and sex.
participants/materials, setting, methods: The acute calamity was at Port Arthur, Tasmania, Australia. On 28 April 1996,
a gunman opened fire on visitors and staff in a tourist cafe. A very stressful 20 h period, ended with 35 people dead and 22 injured. A negative
binomial regression model was used to assess the association between this calamity and pregnancy loss. This loss is evidenced by a shortfall in
the registration of pregnancies that were in their first trimester at the time of the calamity.
main results and the role of chance: We estimated a shortfall of 29.5% or 229 registered pregnancies among those in the first
trimester at the time of the calamity (P , 0.001), in the region surrounding the calamity site. There was no sex effect in this shortfall (P ¼ 0.911).
There was no corresponding shortfall in other parts of Tasmania (P ¼ 0.349).
limitations, reasons for caution: The study is descriptive and cannot produce causal inferences. These first trimester miscarriages are estimated statistically and it is understood that gestational age is an estimate. The use of maternal residential postcodes at birth as a
surrogate for geographic area or space assumes that the mother has not moved into the postcode area after the calamity and before the reporting
of a birth.
wider implications of the findings: The results of this study suggest that calamities bring about significant pregnancy loss affecting both sexes. The methodology presented of inferring conception date from birth date and using this for analysis, provides a more accurate
assessment of first trimester pregnancy losses than raw birth data or sex ratio at birth.
study funding /competing interest(s): No funds were sought or applied for in the conduct of this study. There are no competing interests for any of the authors.
trial registration number: This was not a clinical trial.
Key words: miscarriage / stress / pregnancy / acute calamity / first trimester / conception based / acute / sex ratio
& The Author 2015. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.
For Permissions, please email: [email protected]
2672
Introduction
Maternal stress and anxiety in pregnancy has been shown to be associated with adverse outcomes such as preterm birth and infants with
lower birthweights (Austin and Leader, 2000). This, in turn, has
adverse implications for fetal neurodevelopment and infant outcomes.
Women who experienced major life events such as the death of a
family member are at 1.4 –1.8 times greater risk of preterm birth, with
strongest effects when events occur early in pregnancy (Austin et al.,
2005a,b; Schetter and Tanner, 2012).
Previous studies of natural disasters such as earthquakes and hurricanes have shown that there is an increase in the number of preterm
births and low birthweight infants (Simeona, 2009). The greater the disaster, the greater the impact on the outcome (Glynn et al., 2001). The
earlier that mothers are affected in the second trimester of pregnancy,
the greater the negative impact on preterm deliveries. The duration of
exposure as well as the severity of the event affect the chances of suboptimal pregnancy outcomes.
A study of floods in a village in Southern Poland in 1997 (Neuberg et al.,
1998) reported an increased incidence of spontaneous abortion. Both
that study and a study of the flooding of the Red River in North
Dakota in 1997 (Tong et al., 2011) showed an increase in preterm
birth and low birthweight infants, although there was no increase in the
number of small for gestational age infants in the North Dakota sample.
Xiong et al. (2008) found that birth outcomes among the survivors of
Hurricane Katrina depended on maternal exposure to the disaster with
depressed birthweight being the key result. A study on women who were
pregnant during a major ice storm in Quebec, Canada, found that their
children had poorer intellectual and language functioning at 2 years of
age (King and Laplante, 2005).
Camacho (2007) used the explosion of land mines in Colombia to test
the effects of increased exposure to terrorist attacks on pregnancy outcomes. That study found that the incidence of land mine explosions in
close proximity to the mothers during pregnancy was a significant predictor
of low birthweight infants and preterm deliveries. The effects were especially pronounced, if maternal stress occurred in the first and last trimesters.
Elevated levels of the stress hormones corticotrophin-releasing
hormones (CRH) have been associated with poor regulation of stress
and increased fear behaviour in infants (Davis et al., 2005; Weinstock,
2005). Austin et al. (2005a,b) showed that stressed and anxious
mothers were more likely to have infants whose temperament was classified as difficult.
A study by Catalano and Hartig (2001) on very low birthweight using
Swedish data found an association with ‘communal bereavement’.
Another study by Bruckner et al. (2010) suggests that variations in the
fetal death sex ratio show evidence of ‘communal bereavement’.
Acute calamities such as 9/11 have an extended reach well beyond
emergency workers and witnesses.
A number of calamity studies have analysed the human sex ratio at
birth in survivors of these events. Catalano et al. (2005) examined variations in the sex ratio at birth in a Californian population after 9/11.
A second study examined the same variations in New York (Catalano
et al., 2006). A decline in the human sex ratio at birth was reported.
The methodology of using a sex ratio to determine possible pregnancy
loss carries the assumption that male fetuses are more vulnerable than
female fetuses and leads to differential losses of males. The only interpretation open to these studies is that in the aftermath (up to 9 months
Dean et al.
later) of severely traumatic events, more males were lost than females,
hence the variation in the human sex ratio at birth.
Summary
There are little data on the impact of calamity on early pregnancy loss.
The standard practice of recording pregnancies that survive beyond
20 weeks does not allow for estimating earlier pregnancy losses.
Among the many causes of miscarriage, the effects of exogenous calamities on a pregnant population before the beginning of the mandatory
reporting period are not well understood.
Study objective
The objective of this study was to use hospital data in the NPDC (National
Perinatal Data Collection) to establish whether there were pregnancy
losses as a result of a calamity.
To answer this question, we have developed a conception time-based
methodology, which we believe enables us to determine whether an
acute calamity can be associated with pregnancy losses before the beginning of the mandatory reporting period.
Materials and Methods
Ethics approval was obtained from the Department of Health and Human
Services, Tasmania, as well as from the Australian Institute of Health and
Welfare, Canberra, and The University of New South Wales Ethics Committee, Sydney, New South Wales, Australia.
Tasmanian birth data were collected for the period 1991 – 1997 (both
years inclusively) comprising 40 788 births. The data were obtained from
the NPDC, an all Australian States and Territories’ collection of hospital
birth data which includes all births over 20 weeks gestation or weighing
more than 400 g.
Gestational age in Australia is defined as the duration of pregnancy in
completed weeks calculated from the date of the women’s last menstrual
period and her baby’s date of birth. If the dates are not known then the gestational age of the pregnancy is determined by a first trimester ultrasound
examination.
The birth data set for this study has the following variables for each birth:
birth date, gestational age in weeks, maternal residential postcode and sex.
We transform these data to a conception time-based data set by calculating the conception date for each birth as the birth date minus gestational
age plus 2 weeks. Gestational age is a mandatory reporting item in the
NPDC collection. It is, of its nature, an estimate.
The three variables in the conception time-based model which are used in
the statistical modelling are estimated conception date, maternal residential
postcode and sex. The result is a form of simulation, which we refer to as conception time-based methodology.
We estimated that 40 318 pregnancies were conceived in the years 1991 –
1996 inclusively (Table I). The difference between pregnancies (40 788) and
conceptions (40 318) occurs because pregnancies in the 1991– 1997 study
period were conceived before 1991. Similarly, many pregnancies ending
after 1996 were conceived within the study period.
Assuming the calamity effect was most severe at and around the site of the
calamity in the south eastern corner of Tasmania, maternal residential postcode was used as a geographic variable for division of the state into two
regions, southern and northern (Fig. 1 postcode map). Pregnancies were
aggregated into disjoint 4-week blocks of conception time and classified by
region and sex. The 12-week period 6 February to 28 April 1996 was
taken as the calamity period, i.e. pregnancies conceived during that period
were in their first trimester at the time of the calamity.
2673
Acute calamity impact on early pregnancy losses
Table I Estimated calendar month-specific number of conceptions for all Tasmania for the years 1991– 1996.
Years
1991
1992
1993
1994
1995
1996
Totals
.............................................................................................................................................................................................
Jan
587
646
615
638
617
586
3689
Feb
539
539
577
552
504
444
3155
Mar
576
582
561
575
504
465
3263
Apr
570
548
532
541
571
467
3229
May
583
564
554
525
530
550
3306
Jun
548
572
598
563
543
506
3330
Jul
588
580
561
636
523
542
3430
Aug
594
562
543
540
532
534
3305
Sep
578
532
575
543
519
501
3248
Oct
589
556
542
579
540
560
3366
Nov
562
592
550
603
570
536
3413
Dec
Totals
632
616
625
615
568
528
3584
6946
6889
6833
6910
6521
6219
40 318
Figure 1 A map showing the entire state of Tasmania. The orange sector in the south-east of the state is the limit of the estimated pregnancy losses. In a
direct line, it is 50 km from Hobart to Port Arthur.
The acute calamity
On 28 April 1996, an armed gunman entered the Broad Arrow cafe in the
well-known tourist venue at Port Arthur on the Tasman Peninsula in
Tasmania, Australia (Fig. 1). He immediately killed 12 people in the space
of 15 s (Cox, 1996). In the following hours, he killed a total of 35 people
and injured 22. Many of the dead and wounded were local residents. The
wounded, some seriously, were immediately evacuated to Hobart hospitals.
By nightfall, the gunman was still at large. That night police closed the
venue and peninsula (Fig. 1) and imposed a lockdown over the whole peninsula. Tourists (both domestic and international), paramedics, residents,
2674
Dean et al.
emergency workers and venue staff, trapped by these security precautions,
spent the night in cramped conditions with furniture pushed against doors
and windows of the many buildings at the venue (Scott, 1997). Many of the
emergency staff were from Hobart and environs (shaded area in Fig. 1).
The Tasman Peninsula is very narrow at its’ northern extremity and a
simple blockade was sufficient to keep the gunman and unfortunately everyone else within the peninsula area. The gunman was arrested the following
morning. The shock from these serial murders was widely publicized and
conveyed throughout the world. We use this incident as the calamitous
event.
Statistical analyses
We assessed the potential effect of the calamity on pregnancies within their
first trimester at the time of the calamity. The number of pregnancies in
4-week blocks of conception time, classified by region and sex, was the dependent variable in a multivariate regression model. The predictors were
secular trend (linear) annual cyclic trend (a cosine term), region, sex and a
binary variable indicating if the conception time corresponded to a pregnancy
within its first trimester at the time of the calamity. (A sine term was originally
included, but, as it was not statistically significant, it was excluded from the
model.) Interaction terms were used to assess the effects of calamity and
sex within region. As a Poisson response distribution provided an inadequate
fit to the data, a negative binomial regression model was used.
In regression modelling, the negative binomial response distribution is
parameterized in terms of its mean and a dispersion parameter, and a loglinear model is specified on the mean (McCullagh and Nelder, 1989).
The Akaike information criterion (Lindsey and Jones, 1996) was used for
model selection and quantile residuals (Dunn and Smyth, 1996) for assessment of model fit.
The statistical language R (version 3.1.2) (RCoreTeam, 2014) was used for
analysis. A significance level of 5% was used for statistical testing.
Figure 2 Number of male and female conceptions (4-weekly totals
shown as separate points). The shaded area represents the conception
times of pregnancies calculated to be in the first trimester at the time of
the calamity (solid vertical line).
Table II Negative binomial regression on 4-weekly
conception totals.
Predictor
Results
Figure 2 shows the number of pregnancies conceived (4-weekly totals)
over the study period, with northern and southern regions indicated,
as well as the calamity period. In the southern region, the registration
of pregnancies that were in the first trimester during the calamity
period exhibit an apparent shortfall.
The negative binomial regression model for pregnancies by conception time (Table II) shows a statistically significant effect of calamity in
the southern region (P , 0.001), after correction for linear secular trend
(P , 0.001), annual cyclic seasonality (P , 0.001), region (P , 0.001)
and sex (P , 0.001). There was no significant effect of the calamity in
the northern region (P ¼ 0.349).
A three-way interaction between calamity, region and sex was not significant (P ¼ 0.911), meaning that the effect of the calamity in the southern region was not significantly different for males and females. (As this
effect was not significant, it was omitted from the final model and is
not shown in Table II.)
The incidence rate ratio (IRR) for calamity in the southern region is
0.705. This is a multiplier on the expected number of registered pregnancies and means that, the number of first trimester pregnancies at the time
of the calamity that proceeded to 20 weeks is reduced by 29.5%.
This represents an expected shortfall of 229 pregnancies. When the
(12 weeks) calamity period was broken down into three 4-week
blocks, there were no significant differences in the effect of the calamity
between the first, second and third blocks (P ¼ 0.471). The IRR for sex
(female) is 0.933, implying a sex ratio of 107 males to 100 females (Fig. 2).
Incidence
rate ratio
95% CI
P-value
........................................................................................
Time (linear)
0.985
(0.978, 0.991)
,0.001
Annual seasonality (cosine)
1.034
(1.019, 1.050)
,0.001
Region (South)
0.890
(0.871, 0.909)
,0.001
,0.001
Sex (female)
0.933
(0.914, 0.953)
Calamity (North)
0.964
(0.892, 1.041)
0.349
Calamity (South)
0.705
(0.625, 0.795)
,0.001
The predictor ‘Calamity’ takes the value 1 when conception was in the first trimester at
the time of the calamity.
Discussion
We have demonstrated a substantial loss of first trimester pregnancies
at the time of the calamity, in the region of the calamity. The current
practice of mandatory recording of all births that occur after 20 weeks
gestation does not allow for the reporting of first trimester pregnancy
losses. Our methodology allows for the estimation of shortfalls in pregnancies proceeding to 20 weeks, by taking individual recorded births and
using their estimated gestational age to calculate conception date.
We suggest that the shortfall of 229 first-trimester pregnancies were
miscarriages as a result of calamity from maternal stress and anxiety.
Causes for the shortfall could be: (i) miscarriage as a result of calamity
induced stress, (ii) infertility prior to the calamity (and therefore not
caused by it), or (iii) terminations of pregnancy. There are no reasons
or evidence to suggest that (ii) or (iii) have any validity.
2675
Acute calamity impact on early pregnancy losses
As Popper (1959) points out, the closeness of time (immediate 12
weeks before the calamity) and space (the area immediately around
the calamity site) strengthen the case that the observed association
is causal. Trusted (1979), while endorsing Popper’s maxim, added statistical significance to argue for causation. The size of the shortfall suggests
that the calamity effect was not limited to victims and eye witnesses.
There was no significant sex effect in the shortfall. Male and female
fetuses were affected to the same extent. Our methodology produces
an estimate of total estimated pregnancy loss, thus differing from methodologies which focus on the human sex ratio at birth (Catalano et al.,
2005, 2006). The latter exclude any estimate of female fetal loss, with
an underlying assumption of male fetal vulnerability and female fetal
robustness.
In later studies, hypothesis that populations without firsthand knowledge of people involved in a calamity or are some geographical distance
from a calamity nevertheless show signs of distress (Bruckner et al.,
2010). Intimacy is not the only factor when communicating grief.
It is significant to note that while the Tasman Peninsula is sparsely
populated, Hobart and environs has a population of nearly 200 000
people. This calamity affected directly; hospital staff, doctors, paramedics, police officers, forensic staff, fire fighters and everyone who
entered the Broad Arrow Café that afternoon after the gunman had
left (Scott, 1997).
The pathophysiological mechanism of how stress may cause early
pregnancy loss has been explored by a number of studies. Nepomnaschy
et al. (2006) showed that women with increased levels of urinary cortisol
levels in the first 3 weeks after conception were more likely to miscarry.
In a further publication, Nepomnaschy et al. (2007) explored the endocrine and immunological effects of stress in reproduction.
Women with increased levels of cortisol tend to have lower level of
progesterone which in turn is associated with an increased miscarriage
rate. James (2015) suggested that stress induced higher levels of
adrenal androgens which were an important factor in early miscarriages.
Wainstock et al. (2013) showed women exposed to repeated rocket
attacks were at an increased risk of miscarrying.
The medical and epidemiological literature suggests that the increased
production of neuropeptide CRH during periods of anxiety plays a role in
initiating labour. Women with elevated blood levels of CRH in the
second trimester or early third trimester are at a higher risk of preterm
deliveries (Wadhwa et al., 1998; Hobel et al., 1999; McLean et al.,
1999; Holzman et al., 2001; Hobel et al., 2008; Chen et al., 2010)
Neuroendocrine studies have demonstrated that the acute response
to stress is associated with a rise in cortisol secretion from the adrenal
gland. If the stressor becomes chronic, the hypothalamic –pituitary –
adrenal (HPA) axis may be abnormally reset and chronically raised
CRH and cortisol levels can ensue (Austin et al., 2005b).
Mancuso et al. (2004) showed that maternal stress measured at
28 –30 weeks of gestation correlated significantly with maternal CRH
levels measured at the same time. Women who had higher levels of
measurable CRH at both 18 –20 and 28 –30 weeks delivered earlier
than women with lower CRH levels. Kalantaridou et al. (2010) questioned whether it was the raised level of CRH that led to preterm birth
or whether the raised level was a consequence of the underlying pathophysiology.
It appears that antenatal fetal exposure to small amounts of stress may
be necessary for normal infant development. In some cases, earlier maturity of the fetal HPA axis can be protective and produce some beneficial
effects. For example, preterm infants whose mothers developed chorioamnionitis had a decreased requirement for respiratory support in
the first week of post-natal life (Watterberg et al., 1997).
What is not known at present is the level of maternal stress, in terms of
both amount and duration that may become harmful to the fetus and
consequently the child in the longer term.
Limitations of the study
This methodology allows for estimating pregnancy losses that occurred
before the mandatory reporting period. As such losses are not reported,
we can only offer an estimate based on gestational age and time to
ovulation as noted in the next paragraph.
Calculation of conception date was based on the assumption that
the last menstrual period for each maternal case was 2 weeks before
implantation.
We acknowledge that some women may have irregular or longer
cycles and that reporting of the last menstrual cycle may not be accurate
in all cases.
The use of maternal residential postcodes as a surrogate for geographic area or space assumes that the mother has not moved into the postcode area after the calamity and before the reporting of a birth.
Summary and conclusion
In conclusion, we record the following findings:
(i) We have developed a statistical model which estimates excess first
trimester pregnancy losses at the time of an acute calamity.
This conception time-based methodology illuminates an area of
pregnancy loss not previously identified. Routinely, collected birth
data only records birth losses after 20 weeks gestation.
(ii) A statistically significant shortfall in first trimester pregnancies was
estimated in the geographic region of the calamity at the time of
the calamity. The size of the shortfall suggests that the effect of
the calamity was not limited to victims and eye witnesses.
(iii) No sex effect was observed in the shortfall.
This study is the first to estimate excess first trimester pregnancy losses
after an acute calamity. This model establishes a conception-based methodology which can be utilized to estimate excess pregnancy losses within
a geographical location, by time and sex.
Authors’ roles
R.G.D. conceived original idea and study design, conducted literature
review, performed data and statistical analysis, and drafted and edited
this paper. J.D. provided statistical guidance, examined the efficacy of
the original idea, data selection and transformation from birth to conception, extraction and analysis, editing this manuscript and provided the
postcode Tasmania map. G.Z.H. performed the data analysis and critically reviewed and edited the manuscript. L.R.L. reviewed and drafted
and edited the relevant literature, provided data analysis and critically
reviewed the manuscript, tables and figures.
Funding
No funding was received for this study.
2676
Conflict of interest
None declared.
References
Austin MP, Leader L. Maternal stress and obstetric and infant outcomes:
epidemiological findings and neuroendocrine mechanisms. Aust N Z J
Obstet Gynaecol 2000;40:331 – 337.
Austin M-P, Hadzi-Pavlovic D, Leader L, Saint K, Parker G. Maternal trait
anxiety, depression and life event stress in pregnancy: relationships with
infant temperament. Early Hum Dev 2005a;81:183 – 190.
Austin M-P, Leader LR, Reilly N. Prenatal stress, the hypothalamic –
pituitary – adrenal axis, and fetal and infant neurobehaviour. Early Hum
Dev 2005b;81:917– 926.
Bruckner TA, Catalano R, Ahern J. Male fetal loss in the US following
the terrorist attacks of September 11, 2001. BMC Public Health 2010;
10:273.
Camacho A. Stress and birth outcomes: evidence from terrorist attacks in
Columbia. CEDE Working Paper 2007 – 2008, 2007.
Catalano R, Hartig T. Communal bereavement and the incidence of very
low birthweight in Sweden. J Health Soc Behav 2001;42:333 – 341.
Catalano R, Bruckner T, Gould J, Eskenazi B, Anderson E. Sex ratios in
California following the terrorist attacks of September 11, 2001. Hum
Reprod 2005;20:1221 – 1227.
Catalano R, Bruckner T, Marks AR, Eskenazi B. Exogenous shocks to the
human sex ratio: the case of September 11th, 2001 in New York City.
Hum Reprod 2006;21:3127 – 3131.
Chen Y, Holzman C, Chung H, Senagore P, Talge NM, Siler-Khodr T. Levels
of maternal serum corticotropin-releasing hormone (CRH) at midpregnancy
in relation to maternal characteristics. Psychoneuroendocrinology 2010;
35:820 – 832.
Cox CJ. R vs Bryant Sentencing Comments. Supreme Court of Tasmania,
Hobart, 1996.
Davis E, Glynn LM, Dunkel Schetter C, Hobel C, Chicz-Demet A,
Sandman CA. Corticotropin-releasing hormone during pregnancy is
associated with infant temperament. Dev Neurosci 2005;27:299 – 305.
Dunn PK, Smyth GK. Randomized quantile residuals. J Comput Graph Stat
1996;5:236 – 244.
Glynn L, Wadhwa P, Dunkel-Schetter C, Chicz-DeMet A, Sandman C.
When stress happens matters: effects of earthquake timing on stress
responsivity in pregnancy. Am J Obstet Gynecol 2001;184:637.
Hobel CJ, Dunkel-Schetter C, Roesch SC, Castro LC, Arora CP. Maternal
plasma corticotropin-releasing hormone associated with stress at 20
weeks’ gestation in pregnancies ending in preterm delivery. Am J Obstet
Gynecol 1999;180:S257 – S263.
Hobel C, Goldstein A, Barret E. Psychosocial stress and pregnancy outcome.
Clin Obstet Gynecol 2008;51:333.
Holzman C, Jetton J, Siler-Khodr T, Fisher R, Rip T. Second trimester
corticotropin-releasing hormone levels in relation to preterm delivery
and ethnicity. Obstet Gynecol 2001;97:657 – 663.
James WH. Hypothesis: high levels of maternal adrenal androgens are
a major cause of miscarriage and other forms of reproductive
suboptimality. J Theor Biol 2015;364:316– 320.
Dean et al.
Kalantaridou S, Zoumakis E, Makrigiannakis A, Lavasidis L, Vrekoussis T,
Chrousos G. Corticotropin-releasing hormone, stress and human
reproduction: an update. J Reprod Immunol 2010;85:33 – 39.
King S, Laplante DP. The effects of prenatal maternal stress on children’s
cognitive development: Project Ice Storm. Stress 2005;8:35 – 45.
Lindsey JK, Jones B. Choosing among generalized linear models applied to
medical data. Stat Med 1996;17:59 – 68.
Mancuso RA, Schetter CD, Rini CM, Roesch SC, Hobel CJ. Maternal prenatal
anxiety and corticotropin-releasing hormone associated with timing of
delivery. Psychosom Med 2004;66:762 – 769.
McCullagh P, Nelder A. Generalized Linear Models, 2nd edn. Boca Raton,
USA: Chapman and Hall, 1989.
McLean M, Bisits A, Davies J, Walters W, Hackshaw A, De Voss K, Smith R.
Predicting risk of preterm delivery by second-trimester measurement of
maternal plasma corticotropin-releasing hormone and a-fetoprotein
concentrations. Am J Obstet Gynecol 1999;181:207– 215.
Nepomnaschy PA, Welch KB, McConnell DS, Low BS, Strassmann BI,
England BG. Cortisol levels and very early pregnancy loss in humans.
Proc Natl Acad Sci USA 2006;103:3938– 3942.
Nepomnaschy PA, Sheiner E, Mastorakos G, Arck PC. Stress, immune
function, and women’s reproduction. Ann N Y Acad Sci 2007;1113:
350 – 364.
Neuberg M, Pawłosek W, Lopuszański M, Neuberg J. The analysis of the
course of pregnancy, delivery and postpartum among women touched
by flood disaster in Kotlin Kłodzki in July 1997. Ginekologia polska 1998;
69:866.
Popper K. The Logic of Scientific Discovery. London: Routledge, 1959.
RCoreTeam. R: A language and environment for statistical computing. Vienna,
Austria, 2014.
Schetter CD, Tanner L. Anxiety, depression and stress in pregnancy:
implications for mothers, children, research, and practice. Curr Opin
Psychiatry 2012;25:141 – 148.
Scott M. Port Arthur: A Story of Strength and Courage. Australia: Random House,
1997.
Simeona E. Out of sight, out of mind? The impact of natural disasters on
pregnancy outcomes. CESIFO Working Paper no. 2814, 2009.
Tong VT, Zotti ME, Hsia J. Impact of the Red River catastrophic flood on
women giving birth in North Dakota, 1994 – 2000. Matern Child Health J
2011;15:281 – 288.
Trusted J. The Logic of Scientific Inference: An Introduction. London: The
MacMillan Press Limited, 1979.
Wadhwa PD, Porto M, Garite TJ, Chicz-DeMet A, Sandman CA. Maternal
corticotropin-releasing hormone levels in the early third trimester
predict length of gestation in human pregnancy. Am J Obstet Gynecol
1998;179:1079 – 1085.
Wainstock T, Lerner-Geva L, Glasser S, Shoham-Vardi I, Anteby EY. Prenatal
stress and risk of spontaneous abortion. Psychosom Med 2013;
75:228 – 235.
Watterberg KL, Scott SM, Naeye RL. Chorioamnionitis, cortisol, and acute
lung disease in very low birth weight infants. Pediatrics 1997;99:2.
Weinstock M. The potential influence of maternal stress hormones on
development and mental health of the offspring. Brain Behav Immun
2005;19:296 – 308.
Xiong X, Harville EW, Mattison DR, Elkind-Hirsch K, Pridjian G, Buekens P.
Exposure to Hurricane Katrina, post-traumatic stress disorder and birth
outcomes. Am J Med Sci 2008;336:111.